Hacker News

Top stories

Live mirror
30 storiesupdated just nowView source snapshot
  1. Claude Code now reads AGENTS.md if there is no Claude.md(claude.com ↗)
    42comments
  2. Android 17 is the first since 3.x to add new APIs without releasing to the AOSP(grapheneos.social ↗)
    142comments
  3. Saving another 100TB of RAM(cloudflare.com ↗)
    21comments
  4. Cloudflare Quick Tunnels(cloudflare.com ↗)
    205comments
  5. Xcode 27.1 Beta Release Notes(developer.apple.com ↗)
    50comments
  6. Cache-to-Cache: Direct Semantic Communication Between LLMs (2025)(arxiv.org ↗)
    10comments
  7. Photon-Emission-Guided Laser Fault Injection Enables RP2350 Secure Debug(ledger.com ↗)
    39comments
  8. Show HN: Cactus Needle 3: 8-29MB automation models can match DeepSeek V4 Flash(cactuscompute.com ↗)
    69comments
  9. How to Write with an LLM(sockpuppet.org ↗)
    221comments
  10. OpenJev(openjev.com ↗)
    235comments
  11. The Implications of Linguistic Illegibility for LLM Security(arxiv.org ↗)
    12comments
  12. C++26: Trivial infinite loops are no longer undefined behaviour(sandordargo.com ↗)
    148comments
  13. Our brain evolved from two primitive nervous systems that merged: Study(newscientist.com ↗)
    43comments
  14. The first new cat species discovered in 100 years(nationalgeographic.com ↗)
    24comments
  15. Korea raises data breach fines to 10% of revenue(koreajoongangdaily.com ↗)
    47comments
  16. I vibed a proof of Conway's conjecture(overreacted.io ↗)
    170comments
  17. How SpaceX streamlined the Raptor engine(construction-physics.com ↗)
    15comments
  18. A search-and-inference database from scratch in pure Zig(antfly.io ↗)
    12comments
  19. Minimal Phone 2(minimalcompany.com ↗)
    125comments
  20. From Geometry to Algebra and Back Again: 4000 Years of Papers (2023) [video](youtube.com ↗)
    discuss
  21. Border agents can search cellphones without a warrant or reasonable suspicion(lawandcrime.com ↗)
    105comments
  22. Inside ZCode: Silently uploading your Git history to the cloud(ferstar.org ↗)
    87comments
  23. Cyclomatic Complexity in C#(ndepend.com ↗)
    2comments
  24. Warez: The Infrastructure and Aesthetics of Piracy (2021)(archive.org ↗)
    8comments
  25. Cekura (YC F24) Is Hiring(ycombinator.com ↗)
    discuss
  26. Show HN: Ax-check.com – Can agents use your product?(ax-check.com ↗)
    24comments
  27. US Military had close call after using AI for hallucinated intelligence report(cnn.com ↗)
    244comments
  28. North Korean nuclear test sets off years of earthquakes(science.org ↗)
    144comments
  29. Mathematicians Build Long-Awaited Graph Sandwich(quantamagazine.org ↗)
    15comments
  30. Show HN: Scry, programmable internet search w/ congestion pricing(scry.io ↗)
    16comments

Low-cost ventilator wins Sloan health care prize (2019)

99 pointsby 6y agonews.mit.edu
28 comments
6y agoHN ↗

Anyone has a link to a description of the actual ventilator? I'm an anesthesiologist/intensivist so I'm a specialist in mechanical ventilation. Would be fun to see what they came up with!

6y agoHN ↗

Thanks! Link doesn't resolve for me. I'm outside US...

6y agoHN ↗

I have an oxygen concentrator, a CPAP/APAP machine and a finger oxymeter. Do you maybe have any links to a crash course into how to use them as an improvised non invasive ventilator, more specifically on how to decide what pressure to use, or maybe can suggest some relevant keywords?

6y agoHN ↗

https://pastebin.com/BE2aTQJ4

This Seattle ICU doctor is recommending that other doctors just generally skip positive pressure ventilation and move directly to intubation, because they feel it's inevitable that the patient will shortly need intubation if they currently need positive pressure.

"given the inevitable rapid progression to ETT once resp decompensation begins, we and other hosps, including Wuhan, are doing early intubation. Facemask is fine, but if needing HFNC or NPPV just tube them. They definitely will need a tube anyway, & no point risking the aerosols."

6y agoHN ↗

Relating to coronavirus (COVID-19), if you require a ventilator, your survival chances are already slim.

The median time from illness onset (ie, before admission) to discharge was 22·0 days (IQR 18·0–25·0), whereas the median time to death was 18·5 days (15·0–22·0; table 2). 32 patients required invasive mechanical ventilation, of whom 31 (97%) died. The median time from illness onset to invasive mechanical ventilation was 14·5 days (12·0–19·0). Extracorporeal membrane oxygenation was used in three patients, none of whom survived. Sepsis was the most frequently observed complication, followed by respiratory failure, ARDS, heart failure, and septic shock (table 2). Half of non-survivors experienced a secondary infection, and ventilator-associated pneumonia occurred in ten (31%) of 32 patients requiring invasive mechanical ventilation. The frequency of complications were higher in non-survivors than survivors (table 2)

https://www.thelancet.com/action/showPdf?pii=S0140-6736%2820...

6y agoHN ↗

while true, it is super early in this disease. I'd think outcomes would improve over time as there is more experience, and more medication testing happening right now

6y agoHN ↗

In the WHO China report, they mentioned that improvements in the standard of care caused a big reduction on the death rate. But surprisingly there were no details about those improvements! Basically China already has a lot of experience treating this disease, but I did not see any details about what they learned. I still hope that maybe there is some communication directly between the medical professionals, but probably this is just wishful thinking.

6y agoHN ↗

According to Bruce Aylward, of Canada’s WHO, they [China] had standard treatment memos, that reached the 6th iteration (to his surprise), that each hospital was implementing the very next day, after it was received. [1]. At that point, the containment effort was only seven weeks old, so they were updating the guidelines frequently, and distributing it throughout the country.

I don’t know what it is, as I haven’t seen it myself, but they seemed to have the treatment process down to a science. Or at least, to the best of their abilities, given the fast moving situation.

Maybe they’re sharing these guidelines with Italy now, so perhaps doctors and nurses from Italy can shed some light on them.

[1] https://www.youtube.com/watch?v=-o0q1XMRKYM

6y agoHN ↗

Ventilator availability is limiting our ability to get care to the most people we can.

6y agoHN ↗

True, but at least Italy claims that the main bottleneck in ventilator availability isn't in the machines themselves but in trained doctors and nurses to support patients on ventilators.

6y agoHN ↗

What if people aren’t ventilated early enough? Would that increase mortality?

6y agoHN ↗

Not a doctor but with I've been reading yes! xx% have their lungs drowning in water, and need to be intubated. Death rates differ based on care received, Wuhan was overrun with sick people and a lot died. Percentage wise, a lot less died in other parts of China.

6y agoHN ↗

Arguable.

Needing ventilation (i.e developing ARF/ARDS) is not necessarily the cause of mortality rather suggestive of decompensated sepsis. From the limited literature available the causes of death are multi organ failure secondary to sepsis or heart failure.

Ventilating is essentially just a form of life support and not a treatment itself. Certainly not having the resources or equipment to ventilate someone in respiratory failure would cause mortality but the mere act of being able to ventilate someone does not mean you are going to prevent death, only if the goal is to buy time for other things / the immune system to start working.

The timing of ventilation / early ventilation is studied in other disease processes (probably most relevant is pneumonia) with mixed evidence, and we have therapeutics that work for pneumonia.

6y agoHN ↗

You can't draw this conclusion from the linked study due to how the sample was constructed.

"In this retrospective, multicentre cohort study, we included all adult inpatients (≥18 years old) with laboratory confirmed COVID-19 from Jinyintan Hospital and Wuhan Pulmonary Hospital (Wuhan, China) who had been discharged or had died by Jan 31, 2020."

There might be a bunch of patients who needed ventilation and went on to survive, but had not been discharged before the cut-off date.

6y agoHN ↗

I know nothing about this field, however when I looked up "invasive mechanical ventilation" Wikipedia told me: "Mechanical ventilation is termed "invasive" if it involves any instrument inside the trachea through the mouth, such as an endotracheal tube or the skin, such as a tracheostomy tube. Face or nasal masks are used for non-invasive ventilation in appropriately selected conscious patients. [https://en.wikipedia.org/wiki/Mechanical_ventilation]

The device is pictured here https://pbs.twimg.com/media/ES5Skn-X0AAkn3j?format=jpg, according to another comment, and it has a "Face or nasal mask" and is therefore non-invasive. The gloomy death rate cited in the Lancet article not be relevant to this machine and it might be a genuine lifesaver in the care of COVID-19 patients.

6y agoHN ↗

I would say that your definition is a little too strict. There are devices which rest on the top of the larynx (Laryngeal Mask Airway or LMA). These devices are not within the trachea but can only be used in those under a general anaesthetic and such are usually involved in the category invasive.

In essence, the main difference is invasive requires sedation (or the patient has a tracheostomy which is a bit of a edge case)

6y agoHN ↗

A "face or nasal mask" wouldn't be invasive by your definition, either.

6y agoHN ↗

https://pastebin.com/BE2aTQJ4

Seattle-area ICU doctor gives the impression that if a patient needs non-invasive assistance, that implies they will quickly need invasive assistance:

given the inevitable rapid progression to ETT once resp decompensation begins, we and other hosps, including Wuhan, are doing early intubation. Facemask is fine, but if needing HFNC or NPPV just tube them. They definitely will need a tube anyway, & no point risking the aerosols.

6y agoHN ↗

The idea is not solely and originally Piracha’s.

https://www.techjuice.pk/pakistani-phd-alleges-that-his-rese...

Mohammad Mujeebur Rahman was an acquaintance of mine when he was a grad student and I was an undergrad at Caltech (in the MSA, although I have friends who are closer to him). The authors of this article should think twice about giving Piracha undiluted credit.

6y agoHN ↗

It's a news office press release. Nobody should expect it to be a well-researched piece of independent reporting. Nothing more than marketing.

6y agoHN ↗

This is cool, but if you spend 10 seconds in ICU with a patient needing ventilation you will immediately realise the process can’t be scaled up merely by virtue of having more machines.

This is why the only way to avoid the Italian situation is minimising and spreading out the case load.

6y agoHN ↗

Great! Let’s get China to ramp up the production.